Adaptive large neighborhood search for autonomous electric vehicle scheduling in airport baggage transport service

Abstract

Efficient airport baggage transport plays a vital role in reducing aircraft turnaround time, diminishing potential flight delays, and lowering the operation cost. Although the traditional tug-and-dolly system provides operational flexibility, its scheduling is complex and relies heavily on experts’ experience, leading to a low utilization rate of resources and inefficient transport services. To tackle this problem and improve the sustainability of airport ground handling service, this paper proposes a novel scheduling mode using autonomous electric dollies (AE-Dollies)¹ for airport baggage transport. The scheduling of AE-Dollies is modeled as a Split-Demand Multi-Trip Electric Vehicle Routing Problem (SD-MT-EVRP), which considers rich requirements in practical scenarios. An improved Adaptive Large Neighborhood Search (ALNS) based solution algorithm is developed, which integrates several specially designed removal heuristics and a greedy-based charging station relocation algorithm. Extensive computational experiments are conducted, and results show our method is more effective in improving vehicle utilization than the existing method. Moreover, an experimental case study based on Hong Kong International Airport demonstrates the potential use of our method in real-life scenarios.